The Linux FreeS/WAN project is a funded project. It has the mandate to produce an IPsec implementation for Linux. The ultimate goal of this effort is to provide systems and software to permit citizens for the world to keep all of their Internet traffic private. 1.
Testing networking protocols is often difficult. By definition there is at least one network involved and often several independant systems attached to the network.
With many application layer network protocols (e.g. http) one can cheat - the network is the virtual ``loopback'' device, and multitasking permits both ends of a protocol to run on the same host. It is therefore common to see people doing all sorts of network development using a garden variety notebook.
The situation is not the same for transport and network layer protocols such as IP, TCP and IPsec. These layers of the protocol are more fundamental. They are typically implemented inside a system kernel. This makes development work as difficult as generic kernel work.
If one is to test them on one's notebook or desktop, one risks putting ones own development environment at risk. It is common experience that doing kernel development is much easier with at least two machines - one machine is crashed every ten minutes and the other machine is used as the development host. The split between development and testing is much better understood in embedded system work - the machine under test is often of a totally different type than the development machine. Historically, the machine under test (a VCR or a modem) is incapable of even running a development environment.
Network protocol development work is further complicated by the need to have more than one machine involved.